Below I'm publishing an email I just fired off to a trading friend. He and I converse back and forth as I work through my trading journy. I'm in the middle of making a number of changes to my approach and it helps me to detail them in Emails to a friend, to sort of crystalize what I'm doing.

----------------------------------------------------------
I've re-read that paper on Position Sizing effects from the turtles site and it has quite a bit more meaning to me now. I may be over generalizing here but this is my impression of that paper. Basically the experiment suggests that position sizing is important in keeping the trader alive in the markets long enough to profit in an upwardly biased game. I gather the upward bias was to simulate using trading tools that gave you a statistical advantage. Traders who went bankrupt were risking on average about 22% while those who survived the first round risked (on average) 6.6%. Winning traders risked no more than 6%. 40% of those who DID NOT get a lecture on position sizing went broke. Only 6.3% who did hear the lecture went broke. The only influence they had over the game was position sizing.

I don't think they used enough samples in the study quite frankly but it's still very interesting and may be valid.

The percentages aside- the study appears to demonstrate that position sizing can be important to survival in the market. Those who assume too much risk on individual trades can run themselves out of the market and possibly lose it all. Logic can lead us to the same conclusions.

Missing is any way to use money management to move beyond being defensive and to outperform the markets. The control group and the group who heard the position sizing lecture had the same mean returns. So how do traders outperform? I believe that the answer may lie outside of defensive position sizing but may still be in the realm of money management as opposed to superior TA skills.

As you know, I'm not happy with my YTD results in that I have not really been outperforming the markets. I've been down when they are down and up when they are up and generally near average market returns based on the Wiltshire 5000 or the S&P. I am happy that I did not blow up in my first year of trading on my own but the results have not been worth the effort in my mind.

One thing I may do as a result of reviewing this material, is reduce my position sizing from 10% to 6%. At +8% returns on the year at my present stops, I'm ahead of the S&P-but one years results doesn't mean I'm handling my position sizing well enough to avoid disaster. I think that reducing my position sizing here may add quite a bit of downside protection and I'm most likely going to do it. A drop from 10% in position sizing to 6% is a 40% reduction and should decrease the risk of my doing serious damage to myself and spouse! And I'm playing with enough dollars that I can do that without causing trading fees to be a major concern.

I've mentioned to you that Gary Smiths comments to the Traders Talk board made quite an impression on me. Although he is a very active trader and trades much faster than I, his comments still apply. He indicated that it wasn't superior TA skills that allowed him to prosper so much over the years but the way he worked his winners. The way he dealt with trending positions.

For all the reasons above (and others), this has led me to change my application of TA and combine it with a different money management plan. Basically I'm looking to load up on trenders where my initial position is in the black at my stops. I want to take additional buy signals and add on to my buys to better profit from sectors/stocks/countries that are trending. I'm also using a system to drop stops back at certain logical points to improve the odds of their hanging on in a long running trend but still leave some near tops to get what I can if the run is over. Each buy is it's own position and no additional buys will occur until the initial buy is in the black at the stops. So a 18% position in say Japan would consist of two prior positions that are now profitable at the stops and a new position at a logical (TA indicated) buying point. 3 positions does not mean I'm busting my 6% position sizing rule per trade because no new position is added until prior are profitable and I'll trade no small, illiquid stocks that may be more likely to present me with a disaster. I'll still have to limit total position on each stock in case of catastrophe. What I may do is limit myself to three buys on any trending stock which would be a total of 18%. So that a disaster that destroys stops by double digits percents does not also become a disaster for me.

So I'm messing with my TA and messing with my money management in order to reduce my downside risk and also to improve the bottom line.

You will not outperform the market if you increase your positions from 10 to 16......At best you will perform about the same as the market. If you want to outperform in the market you have to take risk........Have less positions and concentrate on the best.....

>>TP: "If you want to outperform in the market you have to take risk........Have less positions and concentrate on the best".....<<

very good.

ed rader

Paul, I'm not sure you understand what I'm saying and I know I'm missing your point. I understand that trading is risk and for risk you seek reward. But there are ways of trading to reduce the very real risk of blowing up (going broke) before your statistical advantage plays out. A trader with a 90% Win to Loss system that bets it all on each individual trade could easily go broke before making any money. For instance, he puts it all into "X" and then there is an announcement that "x" has lied to the investors and is bust, they have no money. The stock opens at zero, the 90% system just went bust.

My risk is 100% of my trading pile. No matter how I trade position size, I will trade 100% of my trading pile. My goal in reducing position size is to reduce risk on any indiviual unit buy. I further define or reduce my risk through the use of stops. My goal in adding to buys in positions going my way is to have the money management funnel my money to postions that are rewarding me for the risk. Loading up where it is doing me the most good. Working my winners" is something I have not done in the past.

Moving from 10% maximium positions in any stock to a 16% maximum in a trending stock (in stages by individudal units) IS taking more risk on that individual stock.

Actually I'm considering 3-6% maximum positions in any stock. So that would be 18% maximum position in any sector or stock.

SB,

I do not know which Turtle study you were refering to above but the original turtes were trading uncorrelated (or not too correlated) futures.

Since stocks more or less fall and rise with the market you should pay even more attention to diversifying. One can diversify across sectors but one also can diversify techniques: trend/countertrend, pairs, seasonals, etc.
One needs every little edge.

I admire your work and determination to succed and wish you the best.

Flyers

When I lived in Las Vegas, I played blackjack every day. I played basic strategy and counted the cards using Dr. Edward Thorpe's high-low count, the simplest, most effective count going.

If I had a 100 bucks, that meant either 20 five-dollar bets or 50 two-dollar bets. You make the most money the quickest with the five dollar bets, but you can go bust quickly too. You make much less per hour with the two-dollar bets, but your risk of busting out is diminished.

It depended on how I was feeling that night and it depended on how the conditions were: was the dealer going deep into the deck? Was I the only person at the table, thus getting more hands per hour? Could I "surrender" a 16 against an ace?

The main thing is: have a money management plan before you sit down at the table and then stick to it. Don't show up and wing it.

The best advice I ever got about money management was from a book on blackjack: Blackbelt In Blackjack by Arnold Snyder.

I know it sounds crazy to bring it up in relation to trading the markets, but it's basically the same thing. In fact, it IS the same thing. There's no difference between bet size and "maximum position."

My best regards,
John

SB, I too commend you on your perseverance but even more for your willingness to try new approaches when the old ways aren't working as well. I knew two of the turtles and one quite well. Back in the early 90s he always blew my mind when he said a prudent trader has to back up every S&P futures contract by $100,000 at a time when I was backing it up with $8,000. It dawned on me then that apparently it's a much different world in the managed money field where there are 100s of millions of dollars on the line versus what us small fry trade. With a smaller account, it is a lot easier to be a more concentrated trader. While some may disagree, I would contend you should trade a $10,000 account differently than a $100,000 account and a $100,000 account differently than a $1,000,000 account and so on as you go up in account size.

I agree with you on the relevance of position sizing and believe it's the most crucial element in the money/risk management genre. Yet, as some have stated in this thread, I find it best to concentrate my money as much as possible in as few of issues as possible. When I traded mutual funds, if the market was running right, I would have all my trading capital in one fund. And with stocks, there have been times when I have had 25% to 30% of my total capital in one stock and as much as 100% in one stock in one of my retirement trading accounts. The key for me in position sizing is *volatility*. Unlike most traders, I absolutely hate volatility. That was the attraction of mutual funds, they were much less volatile than stocks and hence you could trade them more *aggressively* with larger amounts of capital by quickly adding to your winners. And that's why I like the tight rising channels so much in stocks - you can trade them more *aggressively* with more capital. And that is why I love junk bonds so much - you can really trade them *aggressively* and even add with leverage because of their trend persistency and lack of volatility.

I guess what I am saying is position sizing and adding to winners should be a function of volatility. That's why I have had trouble trading lower priced stocks (which I find much more volatile than higher priced stocks) since I haven't figured out how to size my positions in them and even more so how to add to my winners there.

Thanks everyone for your input. I'm going to take it all into account. Hopefully I'm advancing my understanding and my approach.

Risk-based stops are a great way to go in any case, but especially if you want to increase size in profitable positions. This is probably the only way to trade large-cap stocks - if you don't concentrate, with a small account, you will neve win. They are too efficient and you don't have the capital. On the other hand, concentration doesn't necessarily mean more risk if you are being careful about respecting all your stops for all your tranches.

I am taking the opposite route now though, as I have a small account and am willing to take on more risk. I trade fairly volatile low-priced stocks, using risk-based stops and caps on equity percentage. It's sort of a hit and run style, so I don't really have time to add to positions, even if I wanted to. I have not actually done much backtesting with this regard though, I think I may do some today, but my observations and instinct say adding to these will increase risk-of-ruin, and at the very least make for some really killer drawdowns that would probably reduce profitability.

If you're position trading--which I consider holding things for a half dozen to a dozen and a half weeks--you may want to consider something I call a reflex stop. Frequently, when you get taken out at your stop in something you have good gains in you'll find you give back quite a bit of profits from the most profitable point of the position. If you use a reflex stop, you basically wait for you're reflex sell stop to be hit and then put in a sell limit at an arbitraty retrace level back towards the high. To protect yourself, set a final stop point in case of a waterfall like decline. I've seen so often when posiiton trading that a nasty correction takes people out at undesirable points when momentum type stocks typically tend to retest highs on most time frames. This method is not for everyone because it requires LOTS of discipline and a solid working knowledge of the general market environment. It also adds a little complication to trading--but not much. What it does is keep you from selling at the ST/IT weakest points on a chart. Those points are weak because so many stops are being hit.

Just some thoughts here as I'm currently continuing to develop and refine this methodology and further integrate it into my own position trading.

SB,
You have no idea how much your willingness to post this will mean to a lot of the lurkers here. Some of us "old timers" are going to benefit due to reading and saying, maybe I need to review my own risk management. Risk or money management is the most important aspect of trading.
I am honored that you would share with us, me, your thoughts and concerns.
Best to you and THANKS.
mss

I couldn't agree more with the anology between gambling & trading.
Both must have money management and a system.

Both gambling and trading tend to chop you up most of the time but on those rare times you get a trend going, MAN is it fun.

I love to play casino craps.
As with trading I've noticed that different systems work at different times.

I believe one reason that traders are always looking to "tweak" their system
is that no system works all the time because the market conditions change.
Sometimes it is in a trading range, then it changes into a trend, then chaos.

My favorite strategy is to take partial profits off the table and let the winners run on the house's money.

I enjoyed reading your comments above. I cannot agree with you more.

Very best,
Hitoya

I would recommend to anyone learning the basic principles of probability and statistics in the analysis of the time series --this is what basically TA is all about. If one knows the probability of a given situation statistically in terms of the deviation from the mean at a certain statistical correlation level, then one can evaluate basically much better how likely the trade will end up in their favor.

I evaluate the momentum patterns statistically on the historical data since the 1960s and adjust the size of my positions based on the probability of the trade and the profit potential. The larger the standard deviation among all of the correlating cases, the less money I would commit to the position. Basically, I can commit 100% of my money in one stock if the odds are above 90% with little standard deviation (among all of the similar cases) and the profit potential is above 10% with some hedging, no problem...

The general guidelines in the ARIMA and spectrum analysis should be enough for any trader to get a good understanding about the odds. However, one also needs to be well equipped with the computers for this purpose. Personally, I came up with several momentum and cyclical analysis tools, but they would not be useful, if I didn't do the analysis on these tools to calibrate on the historical data. Anyone can come up with a signal tool, but without the proper calibration to teach to the trader about how and where to use the tool, the trend analysis and the trades can be hardly successful.

Essentially, the definition of the momentum is nothing but the statistical correlation of the trend of a mean with its standard deviation; the faster the standard deviation increases with the moving average, the more likely that there is a strong trend. On the other hand, the seasonal or cyclical patterns are to be extracted using a spectrum analysis; this is what Jim Hurst pioneered first.

Both of these analysis are necessary for the proper analysis of the trends. I don't favor one over the other, it would be an incomplete analysis. If one is very strong, it tends to override the other too. For example, you should not expect to see a major dip right after a high momentum break out (compared to the historical data) just because you have a minor cyclical low coming, the simple trendlines can be very misleading in these situations. Similarly, the larger cyclical signatures, say over 20 weeks in the Hurst model, will be usually seen within the momentum patterns.

As a side note, the EWP actually tries to recover the commonly encountered or known momentum patterns out of the price patterns. Hence, even though it can be considered as an art, the trained eyes can spot many of these patterns quickly, but it is very hard to take the bias out of the human perception because the word "common" means statistically more frequent than the others, BUT, not having less influence just because they are not well known or it can be a combination of several patterns. So, you would need an unbiased momentum and cyclical system to guide you through...

- kisa

Edit: tried to clarify a bit more...

I agree, thats the only way I know also.

It helps even my system...surprisingly.

Here is what I did -

1) enter initial position with risk-based stop, risking no more than 1% of capital, additional 15% equity cap just in case
2) once position profit is half of the original risk-based stop add an additional tranche with same formula for the risk-based stop, then cut the second position in half to ensure that no additional risk is being taken, though we are increasing the percentage of equity beyond 15% in some cases.

I could then go on and add another tranche but if I really liked but it will take some time to balance any few bits of profit I could pick up versus transaction costs once you start adding smaller and smaller sizes.

My system doesn't use price targets so that's the best I can do. Note Sharpe ratio, profit factor and Wealth-Lab score drop because of added exposure.

System 2 - adding additional position.
Long + Short Long Only Short Only Buy & Hold
Starting Capital $100,000.00 $100,000.00 $100,000.00 $100,000.00
Ending Capital $823,701.38 $823,701.38 $100,000.00 $317,780.53
Net Profit $723,701.38 $723,701.38 $0.00 $217,780.53
Net Profit % 723.70% 723.70% 0.00% 217.78%
Annualized Gain % 69.66% 69.66% 0.00% 33.62%
Exposure 72.62% 72.62% 0.00% 99.83%

Number of Trades 1,635 1,635 0 743
Avg Profit/Loss $442.63 $442.63 $0.00 $293.11
Avg Profit/Loss % 3.54% 3.54% 0.00% 221.92%
Avg Bars Held 10.61 10.61 0.00 1,004.00

Winning Trades 950 950 0 512
Winning % 58.10% 58.10% N/A 68.91%
Gross Profit $1,376,904.91 $1,376,904.91 $0.00 $228,984.93
Avg Profit $1,449.37 $1,449.37 $0.00 $447.24
Avg Profit % 12.97% 12.97% 0.00% 340.03%
Avg Bars Held 10.29 10.29 0.00 1,004.00
Max Consecutive 16 16 0 N/A

Losing Trades 685 685 0 231
Losing % 41.90% 41.90% N/A 31.09%
Gross Loss $-653,203.53 $-653,203.53 $0.00 $-11,204.39
Avg Loss $-953.58 $-953.58 $0.00 $-48.50
Avg Loss % -9.52% -9.52% 0.00% -39.85%
Avg Bars Held 11.06 11.06 0.00 1,004.00
Max Consecutive 12 12 0 N/A

Max Drawdown $-78,209.94 $-78,209.94 $0.00 $-47,613.58
Max Drawdown % -28.55% -28.55% 0.00% -37.65%
Max Drawdown Date 10/6/2005 10/6/2005 N/A 8/13/2004

Wealth-Lab Score 68.54 68.54 0.00 21.00
Profit Factor 2.11 2.11 0.00 20.44
Recovery Factor 9.25 9.25 N/A 4.57
Payoff Ratio 1.36 1.36 0.00 8.53
Sharpe Ratio 2.00 2.00 0.00 1.32
Ulcer Index 6.38 6.38 0.00 11.87
Wealth-Lab Error Term 11.52 11.52 0.00 11.37
Wealth-Lab Reward Ratio 6.05 6.05 N/A 2.96
Luck Coefficient 13.90 13.90 0.00 31.09
Pessimistic Rate of Return 1.76 1.76 0.00 16.96
Equity Drop Ratio 0.07 0.07 0.00 0.16


System 1

Long + Short Long Only Short Only Buy & Hold
Starting Capital $100,000.00 $100,000.00 $100,000.00 $100,000.00
Ending Capital $686,187.44 $686,187.44 $100,000.00 $317,780.53
Net Profit $586,187.44 $586,187.44 $0.00 $217,780.53
Net Profit % 586.19% 586.19% 0.00% 217.78%
Annualized Gain % 62.06% 62.06% 0.00% 33.62%
Exposure 64.20% 64.20% 0.00% 99.83%

Number of Trades 1,492 1,492 0 743
Avg Profit/Loss $392.89 $392.89 $0.00 $293.11
Avg Profit/Loss % 3.41% 3.41% 0.00% 221.92%
Avg Bars Held 10.27 10.27 0.00 1,004.00

Winning Trades 878 878 0 512
Winning % 58.85% 58.85% N/A 68.91%
Gross Profit $1,108,761.09 $1,108,761.09 $0.00 $228,984.93
Avg Profit $1,262.83 $1,262.83 $0.00 $447.24
Avg Profit % 12.41% 12.41% 0.00% 340.03%
Avg Bars Held 9.55 9.55 0.00 1,004.00
Max Consecutive 17 17 0 N/A

Losing Trades 614 614 0 231
Losing % 41.15% 41.15% N/A 31.09%
Gross Loss $-522,573.59 $-522,573.59 $0.00 $-11,204.39
Avg Loss $-851.10 $-851.10 $0.00 $-48.50
Avg Loss % -9.47% -9.47% 0.00% -39.85%
Avg Bars Held 11.30 11.30 0.00 1,004.00
Max Consecutive 15 15 0 N/A

Max Drawdown $-52,381.19 $-52,381.19 $0.00 $-47,613.58
Max Drawdown % -26.11% -26.11% 0.00% -37.65%
Max Drawdown Date 7/26/2004 7/26/2004 N/A 8/13/2004

Wealth-Lab Score 71.43 71.43 0.00 21.00
Profit Factor 2.12 2.12 0.00 20.44
Recovery Factor 11.19 11.19 N/A 4.57
Payoff Ratio 1.31 1.31 0.00 8.53
Sharpe Ratio 2.09 2.09 0.00 1.32
Ulcer Index 5.95 5.95 0.00 11.87
Wealth-Lab Error Term 10.11 10.11 0.00 11.37
Wealth-Lab Reward Ratio 6.14 6.14 N/A 2.96
Luck Coefficient 14.52 14.52 0.00 31.09
Pessimistic Rate of Return 1.74 1.74 0.00 16.96
Equity Drop Ratio 0.08 0.08 0.00 0.16

Roger: it's probably a topic for another time, but card-counting at blackjack is not gambling. If you do it correctly, you have an edge over the house. That's not true of the game of craps, which IS gambling.

Anyway, I just mentioned it to make a few points about money management...

You are correct.
It's for that reason that casinos do not welcome card counters and use a shoe full of multiple decks and use shuffle master machines (and serve free drinks ) to thwart the efforts of "evil" counters.
My point is about money management AND the way that various systems work for a while but then fail at other times. The similarities to trading are many including the emotional/psychological effects of winning or losing.

SB, Excellent Post..

I wanted to add something for futures traders.

The standard method obviously is "fixed fractions" ie a certain number of contracts per account size (depending on how far you put your stops)

But all of us know, that our system works better in certain market conditions and you have many more winners than losers. It is the opposite in other conditions..

Linking to something I read sometime ago, which may help some to optimise their methods.

Fixed Ratio

What a great topic, Sentient Being! Many great comments too. I think you are one of the relatively rare people with the patience and emotional control and stability to be a really good trader, i.e. you don't get all crazy like the rest of us. (I don't mean to exclude many people here on FF, either -- more that the general populace doesn't put in the time like you do, getting to know yourself, and getting to know the markets.)

You know what you want in the markets, and you're focused on it, and that's a good thing.

In my opinion, the markets mess with our minds, sometimes, (Roger's different systems working at different times) no matter what, so we better know what we want in advance, even if it's just gambling or playing for action. If we begin without focus, the market will likely change to our detriment just as we're getting one.

Hopefully it's not a linear, one-dimensional axis at work where disciplined trading and increasing account size are at one end, with fun at the other end. I can't add much to what you and the others have said -- I'm a "selective plunger" like some, and more than most.

Patience pays off, and I think that there really are times when it pays to increase your bets.



What's the market going to do at blue point "X"? Most of the time that sucker's gonna keep on going up. You see a lot more patterns like "B" than you do "A." So, maybe we go from 6% to 12% long at that point.


Or, maybe we back up the truck! Yeah, Baby!

Best,

Doug